agentleFS
theyoungastronautstheyoungastronauts/polaris/skills/verify/SKILL.md

verify

Skill4 starsChanged 3 months agoMITView on GitHub

Verify a completed plan phase against the plan in a fresh session; produce a PASS/WARN/FAIL report and commit on pass.

188 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use instinct-export, instinct-import and 9 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • TheKREATORSSkill

    instinct-export

    Export instincts from project/global scope to a file

    Also in 3 projects that use this

  • TheKREATORSSkill

    instinct-import

    Import instincts from file or URL into project/global scope

    Also in 3 projects that use this

  • AviadPSkill

    risk-assess

    Assess the risk and blast radius of a proposed code change before implementing it. Evaluates CI impact, framework-wide vs localized scope, rollback path, and identifies blindspots. Use when user says "what's the risk", "is this safe", "assess the impact", "risk assess", "blast radius", or before applying framework-wide or CI-affecting changes. Do NOT use for code quality review (use code-quality) or plan validation (use design-validator).

    Also in 3 projects that use this

  • diego-tobalinaSkill

    build-fix

    Fix build and compilation errors systematically. Run when build fails.

    Also in 6 projects that use this

  • AaravChadhaSkill

    secure

    Security review with a confidence gate - a finding exists only with a concrete exploit scenario and a high/medium/low confidence rating; everything else goes to a worth-hardening list, never inflated. Sweeps five surfaces: auth gates, secrets hygiene, injection and unsafe sinks, unsafe deserialization/crypto/transport, and LLM tool-use trust boundaries. Reports only, never fixes. Use when the user asks for a security review or to check vulnerabilities, secrets, or auth.

    Also in 3 projects that use this

  • ArvorCoSkill

    specify

    Create feature specification from natural language description. Use when starting a new feature or creating a spec. Triggers on: create spec, specify feature, new feature spec.

    Also in 10 projects that use this

  • +1 more from TheKREATORS

The file

Show its rules and text

We split this skill into the rules it gives an agent

17 rules. 0 are common, 11 are distinctive, and 6 are very specific to this project.

  • 0

    Common

    Most of the 167 edited versions we read keep these.

  • 11

    Distinctive

    This author's own choices. Worth a look.

  • 6

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Look for plan.md in the current directory, then ../plan.md.

    line 13
  2. Distinctive:

    If the user specified a phase number, use that.

    line 17
  3. Distinctive:

    Read the phase's objective and tasks, then systematically check:

    line 21
  4. Distinctive:

    Does it work?: Run the test suite. Report pass/fail counts.

    line 23
  5. Distinctive:

    Is it safe?: Check auth, validation, input sanitization, no data leaks.

    line 25
  6. Distinctive:

    Is it scoped correctly?: No work that belongs in other phases, no missing pieces.

    line 27
  7. Distinctive:

    If the phase produced or modified API endpoints:

    line 31
  8. Distinctive:

    Verify the integration summary exists and matches the actual serializers/views

    line 32
  9. Distinctive:

    Check that request/response shapes in the summary match the code

    line 33
  10. Very specific:

    Check if .claude/context/ exists

    line 39
  11. Very specific:

    Read the Last updated date from the header comment

    line 44
  12. Very specific:

    If the date is still the literal YYYY-MM-DD placeholder, the file is an unfilled template — report it as "Template (run /intel)" rather …

    line 45
  13. Distinctive:

    Check ROUTER.md cross-references: verify each listed file actually exists

    line 49
  14. Very specific:

    If context is stale or missing, suggest running /intel to refresh or /remember to add missing entries.

    line 53
  15. Distinctive:

    If the verdict is PASS or PASS WITH WARNINGS (and the user approves), commit all changes for the phase:

    line 103
  16. Very specific:

    Write a clear commit message following commit-conventions (e.g., feat(auth): implement phase 1 — user model and API)

    line 106
  17. Distinctive:

    PASS WITH WARNINGS: Ask the user whether to commit now or address warnings first.

    line 115

“Common” is measured: kept by at least 6 in 10 of the 167 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Verify a Phase

You are reviewing a completed phase against the implementation plan. You should be running in a fresh session — not the same one that wrote the code.

1. Find the Plan

Look for plan.md in the current directory, then ../plan.md. If not found, ask the user where it is.

2. Identify the Phase

If the user specified a phase number, use that. Otherwise, look at recent git history to determine which phase was just implemented, and confirm with the user.

3. Review Against the Plan

Read the phase's objective and tasks, then systematically check:

  1. Does it work? Run the test suite. Report pass/fail counts.
  2. Does it match the plan? Compare each task's acceptance criteria against what was implemented.
  3. Is it safe? Check auth, validation, input sanitization, no data leaks.
  4. Is it maintainable? Readable code, tests present for key paths, no dead code left behind.
  5. Is it scoped correctly? No work that belongs in other phases, no missing pieces.

4. Check Cross-Repo Contracts (if applicable)

If the phase produced or modified API endpoints:

  • Verify the integration summary exists and matches the actual serializers/views
  • Check that request/response shapes in the summary match the code

5. Check Context Health

Perform a lightweight check on the project's context scaffold. This is not a full drift analysis — just surface-level health.

  1. Check if .claude/context/ exists
    • No: Note "Context scaffold not present" and suggest /intel to generate it. Skip the rest of this step.
    • Yes: Continue.
  2. List each file in .claude/context/ (including patterns/*.md):
    • Read the Last updated date from the header comment
    • If the date is still the literal YYYY-MM-DD placeholder, the file is an unfilled template — report it as "Template (run /intel)" rather than stale
    • Note the line count
    • Flag files not updated in 30+ days as stale
  3. Check ROUTER.md cross-references: verify each listed file actually exists
  4. Summarize findings for the report's Context Health section

If context is stale or missing, suggest running /intel to refresh or /remember to add missing entries.

6. Produce the Report

Write the report to the project root's docs/verification/ folder — one level above the sub-project you're working in. For example, if you're verifying work in ~/prj/my-app/api/, the report goes in ~/prj/my-app/docs/verification/phase-N-[name].md. Create the directory if it doesn't exist. All verification reports go here regardless of which sub-project the phase belongs to.

# Verification Report: Phase N — [Name]

## Summary
[One paragraph: overall assessment]

## Results

| Check | Status | Notes |
|-------|--------|-------|
| Tests pass | PASS/FAIL | [details] |
| Matches plan | PASS/FAIL | [details] |
| Security | PASS/WARN/FAIL | [details] |
| Code quality | PASS/WARN | [details] |
| Scope | PASS/WARN | [details] |
| Integration summary | PASS/FAIL/N/A | [details] |

## Context Health

| File | Last Updated | Lines | Status |
|------|-------------|-------|--------|
| context/architecture.md | YYYY-MM-DD | NN | OK / Stale / Missing |
| context/decisions.md | YYYY-MM-DD | NN | OK / Stale / Missing |
| context/conventions.md | YYYY-MM-DD | NN | OK / Stale / Missing |
| context/ROUTER.md | YYYY-MM-DD | NN | OK / Stale / Missing |

[Any broken cross-references or suggestions]

## Issues

### FAIL (must fix)
- [file:line] Description and suggested fix

### WARN (should review)
- [file:line] Description and concern

## Verdict
PASS / PASS WITH WARNINGS / FAIL

7. Commit (on PASS only)

This step applies when you own the phase's work — a top-level manual review session. If a /autopilot or /orchestrator lead spawned you as the reviewer, do not commit: report your verdict and let the lead commit. The session that owns the loop is the one that commits.

If the verdict is PASS or PASS WITH WARNINGS (and the user approves), commit all changes for the phase:

  • Stage all relevant files
  • Write a clear commit message following commit-conventions (e.g., feat(auth): implement phase 1 — user model and API)
  • Include the verification report in the commit

Do not commit if the verdict is FAIL. Fixes must happen first.

8. Next Steps

Based on the verdict:

  • PASS: Changes committed. "Ready for the next phase. Start a new session and run /execute."
  • PASS WITH WARNINGS: Ask the user whether to commit now or address warnings first.
  • FAIL: List what must be fixed. Suggest fixing in this session, then re-running /verify.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.